time dependent displacement

ffcordi

Hi,

I am trying to run a STA_NON_LINE case but I don't understand quite well how to apply a transient displacement to my model. I have read the manual but I think I don't understand well some translations I made with goolge.

Can I apply a displacement that changes in time by the use of defi_fonc? What is the difference with affe_char_meca_F?

Here is my .comm file. I don't have errors when I run but the results are not the expected. Thus I think it may be something related to the way in which I am applying the loading BC.

----------

The model is mainly a cube (one 3D element) that has rollers on three faces. I pull upwards with a displacement that should increase in time. It's a simple model just to check some concepts on cohesive elements.

Attached file:
coh.comm, 2198kb

jeanpierreaubry

hello

what is wrong in the results?

for a first trial i would do a calculation with less steps L_INST has 10000 calculations steps!

also i would try to replace the imposed displacement by an equivalent (elastic) force

i do not know the 'CZM_LIN_REG relation

otherwise eveything seems ok

what is written in the mess file?

jean pierre aubry

JMB365

ffcordi wrote:
The model is mainly a cube (one 3D element) that has rollers on three faces. I pull upwards with a displacement that should increase in time. It's a simple model just to check some concepts on cohesive elements.

Hello ffcordi,

Although your model is a simple single element cube, could you post the *.med file. I am curious to study your problem. Thanks.

Regards, JMB

AsterO'dactyle

ffcordi wrote:
Hi,

Can I apply a displacement that changes in time by the use of defi_fonc? What is the difference with affe_char_meca_F?

Absolutely no difference: BC can been time function by AFFE_CHAR_MECA_F with 'INST' or AFFE_CHAR_MECA with FONC_MULT

ffcordi wrote:
The model is mainly a cube (one 3D element) that has rollers on three faces. I pull upwards with a displacement that should increase in time. It's a simple model just to check some concepts on cohesive elements.

Be careful that CZM laws produces non-uniqueness force-displacement curve. You should use "PILOTAGE" in STAT_NON_LINE (see the figure in attachment)

Attached file:
CZM.png, 6567kb

ffcordi

Hi!

Thank you for all your help!!

It really help to know that my boundary conditions were right. Thus I took a look to the cohesive parameters and found the problem was there (as AsterO'dactyle said)

I was using so much steps because in the cohesive model I want to capture the elastic (linear response), and displacement control because fracture problems are not always stable under force-control.

Thanks also for the recomendation of the PILOTAGE option! In principle I wanted to check stable systems and decided not to use it. Moreover, I tried and works quite good. A bit of deviation but good enought. I wonder if the pilotage is like a "riks method" or something similar?

I post again the comm file and the mail file (I am using aster alone without salome sorry about that). The cohesive parameters are set to have stable conditions. Notice that I have changed the way in which I define the loading condition, hopefully correct.

Thank you all again, :)

Fernando

Attached file:
cohtest.comm, 2934kb

ffcordi

mail file

Attached file:
cohtest.mail, 917kb

pierre0304

If it's possible, can you send me the .comm with the fonction 'PILOTAGE'

Thank you